Making the Monkey

07 Dec 2017

A short interview where Ryan Glass, lead developer on Downtime Monkey and director of Big Toe Web Design, talks about taking Downtime Monkey from idea to reality.

Ryan Glass, Web Developer

Where did the idea for Downtime Monkey come from?

The idea came in 2016 when we had issues with a server and several of the websites that we managed went down. Once we were aware of the situation it was really easy to sort the problem and we were able to get the websites back up and running in just a few minutes. However, the sites had been down for nearly an hour before we became aware the problem - in fact the first we knew of the issue was when one of our clients called us to let us know that their site had gone down - not good! It was quite embarrassing, a little stressful and at the time I remember thinking, 'this can't happen again!'

So then you decided to develop a website monitoring service?

Not right away - at first I looked for ways to be notified when one of our websites went down.

I asked our hosting provider if they would call us if a website went down (in theory the issue was with them as it was a server problem). Although they were a good quality web host with guaranteed 99.9% uptime the best we could get from them was that if they noticed there was a problem they'd email us. In reality, most web hosts don't monitor the websites that are on their servers and if they detect a server problem they try to fix it before their customers notice.

Next, I looked for third party monitoring services but didn't have much luck. There were several options out there but none did quite what we needed. Some were expensive - so expensive that they didn't show prices on their websites! We found some free options and tried out a bunch: a few just didn't work and others looked like they weren't maintained. Some worked for emails but text messages were either unavailable, unreliable or it wasn't clear what costs were involved in receiving texts. It was often difficult to know the costs of text messages if the monitoring service was overseas.

Are text message alerts a big factor?

Very much so. I like to be able to switch off from emails outside of work: if I'm out to dinner I don't want to have to look at my emails but I need to know ASAP if a bunch of websites have gone down. Also, if I'm out of the city I get a phone signal almost everywhere but 3G is much less widespread. I figured text alerts to my phone were the best option.

So what next?

After drawing a blank I decided to write some scripts, just for us to use in house. The scripts monitored our websites and sent out emails and text messages if a site timed-out or went down. It worked really well and we refined things so that uptime stats were recorded and we could view the uptime of each site.

Then, I thought that if I need this then there are probably others in my position too.

And how long did it take to develop Downtime Monkey?

From the start it has been just under a year - when we took the plunge I expected it to take about three months but it took eight in total. Once development was completed we tested the site for about six weeks in house before going live and then carried out several months of public beta testing which went smoothly.

We added some nice features for Pro users such as being able to customise when an alert gets triggered. We found that some websites (maybe those that use cheap hosting providers or bulky content management systems) can run slowly and that time-outs would trigger alerts quite often. Nobody wants to get 50 emails every day saying that their site has dropped for one minute so we made it possible for users to delay alerts until a site has been confirmed down for a length of time that they specify, say 5 minutes. We also made more detailed stats so that people can view the reason for each downtime - it's useful to be able to see if a downtime was recorded because of a timeout or because of a 404.

We've just started promoting the site and it feels great to get to this stage but I realise that now is not the time to celebrate and have a holiday! There are a few other features that we want to implement but the most important thing now is to get the news out that Downtime Monkey is here and to get as many people as possible to sign up and start using our free service.